FanGraphs Baseball


RSS feed for comments on this post.

  1. I think it’s obvious what’s going on here–clearly taller people have better instincts, and so height is a good proxy for whether a player should be a 2nd baseman or 3rd baseman.

    Comment by Thor — December 18, 2008 @ 12:39 pm

  2. There might be a slight advantage to having a taller 3B. With the majority of batters being right-handed you would expect that more line drives would be hit towards the left side of the infield. A taller 3B would have a better chance at snaring the high line drives down the line and would have more effective “step and dive range”. Conversely, a smaller, quicker athlete might have a better chance of covering the larger range and making the pivot at 2B.

    I don’t have the time to parse the batted ball data to determine if my observations are correct.

    Comment by Nick — December 18, 2008 @ 12:54 pm

  3. The assumption you’re making is the same one that MLB teams are making – that smaller = quicker. Dan Uggla isn’t quick. Jeff Kent isn’t quick. Why did they end up at second base? Because they aren’t tall.

    Comment by Dave Cameron — December 18, 2008 @ 12:56 pm

  4. One would think that all else equal, the player with more range should play 2B and the player with the better arm should play 3B … I’m not sure it doesn’t typically work out that way

    Comment by Glen L — December 18, 2008 @ 12:59 pm

  5. If it worked out that way, then when a 3B moved to 2B, their UZR would go down significantly. It doesn’t.

    Comment by Dave Cameron — December 18, 2008 @ 1:01 pm

  6. Dave – You are right. I did make that assumption regarding 2B. However, I hardly think that Jeff Kent and Dan Uggla typify 2B. Being a Cardinals fan I have no doubt that Scott Rolen could handle SS or 2B. Troy Glaus, on the other hand, would have a very difficult time at either position. I understand your point in the article and I mostly agree with you. However, there are some demands of both positions that require specialized skills, i.e. strong throwing arm for 3B and the agility to make the pivot at 2B.

    Comment by Nick — December 18, 2008 @ 1:04 pm

  7. I think among 3B, you get a fair amount of “good fielding 1B” types, who have good gloves and good arms, but limited mobility. Bigger guys will tend to have less range, but that is not so essential at 3B. A fairly simple way of explaining the differences you’re noting is that the bigger guys, who will tend to hit better, are slower than the smaller guys (who hit worse) that end up at 2B.

    There are plenty of good-fielding 3B, e.g. Lowell, Crede, Chavez, probably Rolen too, who I think would perform worse at 2B or SS than a generic translation would indicate, and the same would apply to some speedy but erratic 2B types.

    Comment by Mike — December 18, 2008 @ 1:09 pm

  8. Interesting post. Makes me wonder if Dustin Pedroia (5’8″) and Mike Lowell (6’4″) could switch :) I think alot of what you said makes sense, good fielders are good fielders. Some positions use defense “tools” more exclusively but if a fielder posseses most of the defensive tools available they could play virtually any position with alittle practice.

    Comment by Derek Rabideau — December 18, 2008 @ 1:23 pm

  9. I asked this at the end of the last thread, but it looks like the conversation has moved up here.

    Is 2B really a more important defensive position than 3B (as is assumed)? If it is, and a team has a great defender at 3B and a mediocre defender at 2B, should the great third baseman move to second to maximize the impact of his defensive abilities? Should Adrian Beltre be playing second?

    Comment by mkd — December 18, 2008 @ 1:24 pm

  10. But my enitre point is that this isn’t really true. For years, we’ve heard players described as SS/2B or 3B/1B, as if those were the matched pairs. However, the 3B/2B spots have far more in common defensively than the 3B/1B spots.

    I’m saying that the entire notion that 3rd baseman and 1st baseman are a matched pair is wrong.

    Comment by Dave Cameron — December 18, 2008 @ 1:26 pm

  11. Dave-

    I wonder if there isn’t some bias in the sample in that the group of players who try to play both 2B and 3B in the same time frame are members of the quick and fast with a strong arm group. Has anyone looked into the pool of players who play both to see if they were above or below average at both positions?

    This should be able to control for the two archetype arguments. If someone were tall and slow but with a strong arm they could probably be a league average or below 3B and if they were fast but without a great arm they could probably be an average or below 2B. Teams would probably not be tempted to try these players at the other position, so these players would be interesting to study if they ended up playing both. Those with good range and arms can play both equally well.

    Either way, it backs up your assertion in your earlier post that the positions are roughly equally difficult defensively and should have the same Replacement Level correction. It may just be that the guys who are good enough to stick at 2B based on their speed but weaker arm are fundamentally different from the taller/slower guys who can stick at 3B because of their strong throwing arms.

    Then again, if we find out that below average fielders at 2B are equally below average at 3B and vice versa, then it backs up your assertion that teams are placing players based on pre-conceived notions rather than defensive abilities.

    Comment by CMC_Stags — December 18, 2008 @ 1:27 pm

  12. When allocating defensive resources, would it make more sense to put the better fielder at 3B because more balls are hit to the left side, or is that taken into account when calculating the metric?

    Comment by Nathan — December 18, 2008 @ 1:41 pm

  13. I think it makes sense that height is an advantage, and speed/agility less of an advantage at 3B. They generally play much closer to the batter, and don’t have as much time to do anything before they have to make a play.

    Speed seems like it would be better utilized at 2B, where it takes longer for the ball to get to the fielder.

    Comment by Ryan — December 18, 2008 @ 1:56 pm

  14. No one is saying that speed isn’t more important at second base, or that reaction times and arm strength aren’t more important at third base. We’re all in agreement on that.

    What I’m saying, though, is that height isn’t as good of a proxy for these things as MLB teams may believe. There are slow, short guys playing second base, and fast, tall guys playing third base. Why?

    Comment by Dave Cameron — December 18, 2008 @ 1:58 pm

  15. It’s not so much about the total amount of balls hit into an area as it is about the plays on the margins.

    Comment by Dave Cameron — December 18, 2008 @ 1:59 pm

  16. I’ll be looking at the selection bias aspect of all this shortly.

    Comment by Dave Cameron — December 18, 2008 @ 1:59 pm

  17. Uh, Dave … you don’t seem to be especially familiar with Ellis or Chavez. First of all, yes, both were failed shortstops — but IIRC, the last time Chavez played short in anything other than an emergency-sub situation was high school. Ellis came up through the Royals system (so it wasn’t the A’s who determined by his body type that he should play 2B) playing both SS and 2B about equally (so the Royals didn’t sort him to 2B), and when he was acquired by the A’s, they had Miguel Tejada playing SS (so the A’s only sorted him to 2B out of roster considerations). What’s more, Ellis’s throwing shoulder was destroyed by a collision with Bobby Crosby the year after he was acquired; he’s made a remarkable non-surgical comeback, but his shoulder is no longer capable of making the throws from SS or 3B. What’s more, yes, Chavez is taller — but he’s also put on a lot of weight since coming up to the majors–and as a third baseman; and I suspect that weight profile is similar across a lot of third basemen (of course, part of the sorting you hypothesize could be selecting for guys with frames that can support more body mass).

    The overall pattern you identify may be true, but (a) it’s more complicated than you describe, and (b) Chavez and Ellis aren’t the best tandem to illustrate it.

    Comment by monkeyball — December 18, 2008 @ 2:03 pm

  18. Taller, bigger guys with longer arms can throw further and harder. They play 3b. This doesn’t seem like rocket science…

    Comment by magicman — December 18, 2008 @ 2:05 pm

  19. I understand that, but wouldn’t there likely be more borderline plays if there were more total plays?

    Comment by Nathan — December 18, 2008 @ 2:25 pm

  20. You completely ignore release which is much more important at 2B than 3B

    Comment by Rob — December 18, 2008 @ 2:38 pm

  21. Right – it’s more likely that I’m completely ignorant of the history of those two players than that I might have a valid point.

    Thanks for stopping by.

    Comment by Dave Cameron — December 18, 2008 @ 2:41 pm

  22. Then you’re missing the point.

    Comment by Dave Cameron — December 18, 2008 @ 2:41 pm

  23. Average release rating for 2B in 2008, per the Fans Scouting report: 54
    Average release rating for 3B in 2008, per the Fans Scouting report: 51

    Comment by Dave Cameron — December 18, 2008 @ 2:44 pm

  24. Assuming a player doesn’t have time to run, a taller player has more range because he can reach/dive significantly farther than a shorter player.

    Although I don’t really know what the quantifiable difference is between how long each position has to make a play, it seems like there are a lot more plays at 3B where there isn’t much time to move, compared to 2B.

    Comment by Ryan — December 18, 2008 @ 2:44 pm

  25. It’s also not true … there are almost twice as many balls that go to 2b as 1b. 801 per team to second and 447 to third.

    Comment by devo — December 18, 2008 @ 2:55 pm

  26. Yeah, I agree. Tim Lincecum is 5’10” and he throws a lot slower than Doug Davis who is 6’4″.

    Oh wait.

    Comment by Kyle Boddy — December 18, 2008 @ 3:01 pm

  27. You haven’t proven that taller players have a slower release.

    Comment by Kyle Boddy — December 18, 2008 @ 3:02 pm

  28. I think the logic goes as follows, in no particular order:

    2B’s get hurt (in the field) a lot more than 3B’s.

    Ergo, put your better hitters at 3B.

    Better (or more valuable) hitters hit for power and are generally bigger.

    2B’s have to cover more ground than 3B’s because they play in the middle of the field. Something that inherently favors smaller players.

    And, of course, the old guys say so.

    Comment by Jack F — December 18, 2008 @ 3:14 pm

  29. Your completely right. I obviously wasn’t thinking.

    Comment by Nathan — December 18, 2008 @ 3:33 pm

  30. This is a fascinating discussion, and I’m glad it’s happening in more places now.

    We all agree that you need a strong arm at 3B, and you need more speed at 2B. Do we necessarily agree that being taller, in and of itself, is important at 3B? All other things equal, sure. But, what if the guy 3 inches shorter has a stronger arm and weaker legs than the taller guy? Then what? Who plays 3B and who plays 2b?

    It would seem that we should expect SOME tall weak-armed fast-legged players at 2B and SOME short strong-armed slow-legged players at 3B. After all, the player’s height is a secondary or tertiary consideration.

    But, based on the players we actually see there, the height is a primary requirement. And, since we agree that we don’t really need a tall 3B, then the reasoning for it is that the teams use the height as a proxy. That even the tall guy can run down 90 feet in 4 seconds while the short guy does it in 4.2, then the team bias is to “mentally” add 0.2 seconds to the tall guy and subtract 0.2 to the short guy. (Numbers for illustration only. Illustration only made to speak in numbers instead of adjectives.) Similarly for 3B and strength.

    The larger point is that we can’t necessarily accept that the teams have been able to find the proper equilibrium point between offense and defense at 2B/3B, especially since players are humans, and are not necessarily very accepting of moving around and learning and relearning positions, and being traded from team to team, so that both sides of the aquarium have exactly the same amount of water.

    Comment by TangoTiger — December 18, 2008 @ 3:38 pm

  31. Yes, but not all short guys are quicker. Not all tall guys are slower. And I haven’t any evidence that 2b get hurt more often. If that’s true, that could be a valid point, but I doubt that goes into the thinking at all.

    Way back when, guys like AROD, Cal, etc.. would never have been tried out at SS. They would have been 3b immediately. So I guess the thinking on this has come around some, but it still has a long way to go.

    Comment by lookatthosetwins — December 18, 2008 @ 3:39 pm

  32. I don’t think he’s saying either that you are completely ignorant or that you don’t have a valid point. What he’s saying is, Ellis is not a worthwhile data point because, due to his injury, he is incapable of making the 3b/SS throw at this point. If not for that injury, he would very likely be playing SS for the A’s right now. The decision to place him at second was out of necessity. The team had established starters at 3b and SS (Chavez and Tejada) and the veterans they had around at 2b (Velarde and Menechino) were in the process of proving themselves to no longer be capable MLBers. Over the course of the season, he started a handful of games at 3b as well. He didn’t play much SS, though, since Tejada was doing his best Cal Ripken Jr. impersonation.

    Neither he nor I are arguing with the point of the article, just the example.

    Comment by devo — December 18, 2008 @ 3:40 pm

  33. Theoretical question: if Ichiro were an infielder, but had Frank Thomas’ legs, would we 100% see him at 3B, or is there a 50/50 chance he might be at 2B? What if he had Mike Lowell or Justin Morneau legs (below average, but not horrible)? Any chance he’d be a 2B?

    Q2: If Lastings Milledge or Carl Crawford (fast guys, below average, but not terrible arms) were 6’4″ RH, would we 100% see him at 2B, or is there even like a 25% chance they might be at 3B?

    I think Dave makes a more than fair point that the height-bias is strong enough as to strongly counteract the leg-talent and arm-talent biases.

    Comment by TangoTiger — December 18, 2008 @ 4:00 pm

  34. this was also my main concern with the conclusions- that the group you were analyzing had inherent linearity. I’m most curious about those off this trend (the slow + strong arm, and the fast + weak arm that have played both positions)

    Comment by Tim_the_Beaver — December 18, 2008 @ 4:11 pm

  35. on the other hand, if there are only two 6’3″ middle infielders (H. Ramirez and A. Ramirez), and both are below average at their positions (-1.1 at SS and -9.2 at 2B, 2008 UZR/150, respectively)…

    i know it’s only two guys, and i don’t doubt that some taller/bigger players who could handle 2B get pre-sorted to third based on size/appearance. still, it seems logical to me that being shorter/slighter might be advantageous to playing middle infield positions.

    anyway, i’m out of my league here. thanks for the discussion.

    Comment by patrickc — December 18, 2008 @ 4:18 pm

  36. If Ichiro had Frank Thomas’ legs, I’m not sure that he’s an MLBer … second baseman or otherwise …

    Comment by devo — December 18, 2008 @ 4:22 pm

  37. Presumably, Ichiro with Frank Thomas’ legs would be a lousy 2B. Someone might play him there, but presumably this player would be better at 3B. Point taken, although are there any correlative examples among actual MLB 2B? Or are we just assuming that there must be examples because 3B are taller than 2B?

    Lastings Milledge at 6’4″…well, this isn’t as useful, because at 6’4″ he may have the speed but not the quickness and agility to play 2B. If he’s 6’4″ and moves like Chase Utley, and you’re playing him at 3B, you’re probably wasting his agility and suffering his below-average arm. But again, who are the 3B who should be 2B?

    Comment by patrickc — December 18, 2008 @ 4:28 pm

  38. Why do you doubt that 2B get hurt more often goes into the thinking at all? I don’t know whether it’s true or not, but the standard rap is that middle infielders take a lot of abuse from runners sliding into second. Further, because they have to cover more ground, there may (or may not) be more risk of pulled muscles, twisted knees and ankles, etc. I don’t have any research either, but I don’t understand why you want to write this explanation off.

    Also, I’ve been thinking about Ripken and A-Rod. Who are the above average, “tall” SS who are not also HOF talents? Not saying there aren’t any, but asking who they are…

    Comment by patrickc — December 18, 2008 @ 4:48 pm

  39. I don’t know exactly if 2B get injured more often than other positions, but they do tend to play fewer games, from what I’ve found. It’s entirely likely, in my mind, that this is because they get injured more. Of course, maybe they play fewer games or get injured more because they’re smaller players… (and so on and so on)

    A couple of weeks back, Joe Posnanski was talking about how the various positions were reflected in the HOF, and someone brought up the observation that 2B seem to play fewer games than other positions, and thus are unable to accumulate enough of the counting stats to make certain voters notice them.

    I did a quick survey of 20th century players who had played more than 1000 games at any given position, and calculated the average games those players played over their career (including all positions). I also did this for players with 1200, 1300, and 1400 games played at a given position.

    It holds pretty steady that long-time 2B play fewer games in their career than other positions, though the numbers change some as we limit the group more and more (ie, as we get into the really long lasting players)

    You can see the numbers for each position at this link (be sure to read the comments, as I include more data down there):

    (there was also some good discussion about this point on the forums at Bill James Online)

    Comment by lar — December 18, 2008 @ 5:02 pm

  40. I would imagine that height tends to correlate fairly well with arm strength, so that colors the impressions of the people making these decisions.

    If we have a short guy and a tall guy with exactly the same defensive skills (including arm strength and accuracy), maybe the height discrepancy leads to the notion that the taller guy has a better arm, even if it’s not true in this case. So while the decision to put the tall guy at third in this scenario isn’t based on a real difference in skill, but an idea that is generally true.

    That’s about the best I can come up with.

    Comment by Teej — December 18, 2008 @ 6:29 pm

  41. This is a great discussion. I keep seeking good answers and not finding them. So, as I proceed to offer explanations, bear in mind that I’m not convinced of any of them.

    What’s wrong with the following explanation: guys with strong arms are more likely to have power bats? (Note, the converse needn’t be true.) Now, it might be that height is a common cause of each, but the reason we actually find better bats at 3B would be because better arms tend to make better bats, not because people are specifically selecting for size.

    Comment by philosofool — December 18, 2008 @ 8:50 pm

  42. I didn’t say that there were a ton of large shortstops out there, all i said was that there are some, and there basically weren’t any back in the day. These guys never would have been considered to play shortstop, whether they could or not. This obviously has something to do with the different health and conditioning that we have now, but I think it’s at least a small change in the thinking of scouts and coaches.

    Comment by lookatthosetwins — December 18, 2008 @ 9:59 pm

  43. I like the explanation about range being needed for 2B and not really for 3B.

    Dave makes the argument that guys UZR’s dont really drop when they move from second to third base. This could be because of a small sample size, but it is DEFINITELY affected by the huge selection bias that only certain players are picked to play both 2B and 3B positions.

    Comment by Dave — December 19, 2008 @ 12:19 am

  44. My bad, it looks like those selection issues are being addressed tomorrow

    Comment by Dave — December 19, 2008 @ 1:27 am

  45. Perhaps it’s a comfort/injury thing (though not the injury rate explanation suggested by Jack F above, which I like). This is just wild speculation, but I’ll share my thought:

    Pre-pitch positioning tends to get infielders into a bit of a crouch, ready to spring into action. At 2b, due to covering more ground, the player often has to get a bit upright, run to the ball, and then get back down to make the play. This could be a strain on a taller player’s back even if it doesn’t affect the amount of plays he is able to make. A shorter guy might not feel the effects as much.

    At 3rd, the player gets low before the pitch and usually just reacts from that position to the ball without fully coming to a more upright running position.

    Comment by david h — December 19, 2008 @ 1:51 am

  46. I think there’s an important issue missing here. Most of the guys who have seen a lot of time at both second and third have played a lot more at 2B in their careers (not Iwamura, but certainly most of the guys who have done this). Mostly, they’re utility players who played mostly at 2B and SS in the minors, and are out of position at third. This is because of the perception that 3B is the easiest position, meaning that teams are reluctant to move 3Bs to 2B. Assuming that experience matters, and also that the relative importance of different parts of the skill set matters, I would argue that it is quite likely that it is harder to find a good 2B who is also a good hitter than it is at 3B, and that wholesale conversion of marginal 3Bs into 2Bs isn’t going to work.

    Comment by Tom — January 27, 2009 @ 9:03 am

  47. Perhaps it is an issue of pure durability. The average 2B (min. 1000 innings) over the past 3 seasons started 277.3 games and played in 290.1. The average 3B (min. 1000 innings) over the past 3 seasons, however, started 263.6 games and played in 277.1. Despite the frailties of Ian Kinsler, Aaron Hill, Chase Utley and Rickie Weeks, 2Bs, on average, saw more time on the field than did 3Bs. Then again, 3Bs have had their share of injured played over the past 3 seasons in Joe Crede, Adrian Beltre, Mike Lowell, Aramis Ramirez and Troy Glaus.

    It’s surely not a distribution based on “glove or bat.” While MLB 3B hitters (min. 500 PA) over the past 3 seasons hit .270/.341/.438 (.779 OPS) and 2B hitters (min. 500 PA) over the same 3 season sample size hit .278/.341/.416 (.757 OPS), the average 2B (min. 1000 innings) had only a +0.6 UZR/150 compared to a +0.8 UZR/150 for 3B (min. 1000 innings).

    Any other postulations as to the reason for the distribution? Is it, as you suggest, something cosmetic, or is there something not quantified in the data at play? Does playing 2B, which requires a player to “go both ways” in making outs, require more althletic, healthier types at the helm than does 3B, which generally only requires a player to go to his left in making 80% of his plays? Does a stronger bat equivocate a stronger throwing arm?

    Comment by David MVP Eckstein — February 24, 2010 @ 1:13 am

  48. Smaller guys tend to be more agile. Turning a double-play at 2B requires that agility.

    Comment by M.Twain — February 24, 2010 @ 11:59 pm

Leave a comment

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>


Close this window.

3.386 Powered by WordPress